Saturday, 18 June 2016

Oppo F1 - HDR Mode

I remain a little unclear as to if this mode is worth a rats arse or not.

As I explore it more I'm learning to "ask the right questions" of my photographs to see.

This morning I took some shots which suggested that it didn't really do much, but this after noon I took some shots which suggested that it did.

The answer seems to lay in "how contrasty is the lighting"

So this afternoon while it was overcast I took this shot off the balcony:

and then (wasn't using the tripod) this shot engaging the HDR setting:

The HDR shot actually seems to just get around the cameras dreadfully bad habit of making the shadows look like you'd scared a squid, while doing little to the highlights. As one who's used HDR a lot myself this is most definately NOT taking multiple shots and combining them into a HDR. So its more or less just doing a better job of tonemapping the initial shot into something like a decent photograph.

I confirmed the same again in my bedroom taking a shot of the clothes in my wardrobe (also in dim light) and the HDR setting was better (on the right).

Again, whites nearly the same (lifted perhaps a little) and the reflection on the floor from the door shows curtain shadow better on the right, where its not just "nuclear white"

... The left shot seems to show more shadow noise, so if it had simply dropped exposure then you'd expect that the lower exposure would be noiser (which is the right), so perhaps it is actually giving a bit more than just holding the highlights and beefing up the shadows.

From the EXIF data:
ExposureTime - 1/17 seconds --- 1/20 seconds
ISOSpeedRatings - 1155 --- 800

so it chose a lower ISO on the "HDR" image (which would tend to darken the image, but improve noise) but a shorter exposure on the NORMAL image (which would also tend to darken the image, were it not balanced by the raise in ISO).

That could be part of the algorithm or it could just be that it "picked differently" due to slightly different composition ... I'd need to ask that question in my next test by fixing the camera and fixing the same "metering point" to remove that from the equation.

So you see that each time you take a test shot, the analysis of that outcome guides your exploration for the next iteration.

next I'd like to look at the HDR test from early this morning, when I saw clouds low over our mountain range here being illuminated by the newly rising sun.

again, that squid has squirted its all into the shadows ... and escaped with the shadow details!

So with HDR engaged:

we see the much better preservation of the shadows, without destruction of the highlights (which simply increasing exposure would cause). Seems to fit the pattern.

However what could I have got if I'd taken RAW and processed it with Snapseed on the phone (at the expense of a few moments, which could be done at any time later)?

Well cunningly I prepared for this at the time of taking the first photograph above by: setting RAW capture (which also captures the standard shot at the same instant, writing both the card).

So when I "developed" that RAW file I applied two simple tweaks:
  • pushed the exposure up a little
  • pulled the highlights down (which prevents the clouds from going nuclear)
and that gave a much more dramatic version, capturing the highlights, keeping the shadows and actually rendered the upper altitude clouds nicer against the sky - something the HDR mode failed at.

To my mind if you are in doubt at all, then engage the HDR setting ... I actually think that its what the camera should default to. I suspect even that the HDR setting is more or less what a decent algorithm for writing the JPG should produce, and they only have the crummy standard one to make the HDR appear to do anything.

Of course (as always) if you are even thinking that this is going to be a keeper ... engage RAW


No comments: